Self-Adaption FOPID Controller Design for Internal Temperature in Roller System Based on Data-Driven Technique

In this paper, we desire to update the traditional PID controller of vertical roller mill PID controller to have a better performance in temperature adjustment. The open extent of cold blast sliding valve is our controlled device, and we design a fractional order proportion integration differentiation (FOPID) controller with Particle Swarm Optimization (PSO) algorithm to control it. This algorithm is an assistant tool to achieve online updating the parameters of the FOPID controller. For starters, the pseudo-partial derivative (PPD) parameters are estimated dynamically by the state observer. Secondly, three parameters of integer order PID controller,Kp,Ki and Kd, are determined by PPD parameter observer and gradient descent algorithm. Afterwards, PSO algorithm is used to adjust and correct μ, λ and online update them. Finally, the simulation results show that the controller has a good impact for the outlet temperature control, which is better than the traditional PID controller.